Environ. Sci. Technol. lQQ2,26, 2207-2213
Quantification of Ambient Nitrodibenropyranones: Further Evidence for Atmospheric Mutagen Formation Detiev Helmlg,+ Jos6 L6pez-Cancio,%Janet Amy,
*Is
William P. Harger, and Roger Atkinson§
Statewide Air Pollution Research Center, University of California, Riverside, California 9252 1 ~
~
2- and 4-nitro-GH-dibenzo[b,d]pyran-6-ones were quantified in ambient air samples collected in southern California and in the NIST SRM 1649 urban dust collected in Washington, DC. Lower concentrations of these nitrodibenzopyranones were observed in the NIST SRM 1650 diesel particulate material, and comparison of the concentrations of nitrodibenzopyranones, nitropyrenes, and nitrofluoranthenes in the SRMs 1649 and 1650 supports the conclusion that the nitrodibenzopyranones are formed in the atmosphere. 2-Nitrodibenzopyranone was determined to be a significant contributor of ambient air particle- and gas-phase mutagenicity, as assayed with the microsuspension modification of the standard Ames Salmonella plate incorporation test using strain TA98, without activation. Formation yields of the 2- and 4-nitrodibenzopyranones observed in environmental chamber experiments of the OH radical-initiated reaction of phenanthrene are compared to those derived from ambient air observations.
0 nitro-6H-dibenzo[b,d]pyranS-one (11)from the gas-phase OH radical-initiated reaction of phenanthrene (20, 21). 2-Nitrodibenzopyranone was shown to be a strong direct-acting mutagen in the microsuspension modification (22,23)of the standard Ames plate incorporation test, with a response of 240 000 revertants pg-l (20,21,24),and accounted for the majority of the observed direct-acting mutagenicity of the products collected from gas-phase OH radical-initiated reactions of phenanthrene (21). Furthermore, 2- and 4-nitrodibenzopyranones were observed in a Riverside, CA, ambient particulate extract and 2nitrodibenzopyranone accounted for 20% of the total activity of this crude extract (21). In this work, we have conducted quantitative analyses of the 2- and 4nitrodibenzopyranones present in particle(collected on Teflon-impregnated glass fiber filters) and gas-phase (collected on polyurethane foam plugs) samples collected in southern California. The National Institute of Standards and Technology (NIST) standard reference materials (SRMs) 1649 (urban dust collected in Washington, DC) and 1650 (diesel particulate matter) have also been analyzed for these PAH derivatives. In addition, the formation of 2- and 4-nitrodibenzopyranones and related compounds from the gas-phase OH radical-initiated reactions of phenanthrene, GH-dibenzo[b,d]pyran-6-one and 2,2’-diformylbiphenyl has been investigated in an environmental chamber. The implications of these ambient and environmental chamber data for ambient air mutagenicity are discussed. N
Introduction
It is well-known that extracts of collected ambient air particles are carcinogenic ( 1 ) and are also strongly mutagenic in the absence of microsomal activation (2-5). This direct-acting mutagenicity is not due to the presence of polycyclic aromatic hydrocarbons (PAHs), since they require microsomal activation for expression of their mutagenic activity. However, it has been shown that the nitroarenes, many of which are strong direct-acting mutagens (6), contribute up to -10% of the measured particulate-phase direct-acting mutagenicity of ambient air (7,8). Two important particle-phase nitroarenes observed in ambient air are 2-nitrofluoranthrene and 2-nitropyrene, believed to be formed in the atmosphere from the radical-initiated reactions of gas-phase fluoranthene and pyrene, respectively ( S 1 4 ) . The hydroxyl radical reaction is the dominant atmospheric loss process for the gas-phase PAHs (121,and although the nitroarenes are formed only in low yield (55%))these OH radical-initiated reactions appear to be the major source of atmospherically formed nitroarenes (8, 10, 11, 13-16). The majority of the direct-acting mutagenicity of extra& of collected ambient air particles is from compounds more polar than those in the nitroarene-containing fractions ( 1 7-20). Recently, we employed bioassay-directed chemical analysis of these more polar fractions (20) to identify 2-nitro-6H-dibenzo[b,d]pyran-6-one (I) and 4‘Present address: National Center for Atmospheric Chemistry, Boulder, CO 80307-3000. t Present address: Secci6n de Quimica Analkica, Universidad de Las Palmas, Canary Islands, Spain. AlBo Department of Soil and Environmental Sciences, University of California, Riverside, CA 92521. 0013-936X/92/0926-2207$03.00/0
Experimental Section
Analyses of Standard Reference Materials and Ambient Air Samples. Ambient air samples were collected using Teflon-impregnated glass fiber (TIGF) filters and polyurethane foam (PUF) plugs (8,12,25).The ambient particulate samples were collected on TIGF filters (Pallflex T60A20) using high-volume samplers equipped with 1O-Nm size-selective inlets (General Metal Works Inc.) at calibrated flow rates of 1.1m3min-l (40 SCFM). These samples were collected at Claremont, CA (Harvey Mudd College campus), and Long Beach, CA (Long Beach City College campus), during the 1987 South Coast Air Quality Study (26). Ambient atmospheric particulate samples were also collected at the University of California, Riverside campus, in September and October 1991, during air pollution episodes when the measured maximum daytime
@ 1992 American Chemical Society
Environ. Sci. Technol., Vol. 26, No. 11, 1992 2207
ozone mixing ratios were 170-230 parts per billion. The NIST SRM 1649 urban dust is a composite sample coll e d in the Washington, DC, area over a 12-month period using a specially designed baghouse (27). The SRM 1650 is a sample of diesel particulate material, representative of heavy-duty engine particulate emissions (27). The gas-phase samples were collected at Claremont during the August 1987 South Coast Air Quality Study and at Riverside during September and October 1991 on preextracted (dichloromethane and methanol) PUF plugs (9 cm diameter X 5 cm thickness) downstream of TIGF filters using a high-volume sampler operated at -0.8 m3 min-' (28 SCFM). After sample collection, the filters and PUF plugs were first spiked with appropriate amounts of 3-nitrodibenzopyranone, to serve as an internal standard, and then Soxhlet-extracted with 200 mL of dichloromethane for 24 h. It was confirmed from control analyses of nonspiked filter samples that this internal standard was not present at significant levels. For the SRM 1649 urban dust, 200-mg sample aliquota were Soxhlet-extracted in 2-mm cellulose thimbles (Whatman). The extracts were reduced to -3mL volumes with a rotary evaporator, filtered through 0.45-pm Acrodisc Teflon filters, and then reduced to 200-pL volumes under a flow of dry nitrogen. The concentrates were separated into fractions of increasing polarity by high-performance liquid chromatography (HPLC) using a semipreparative Regis Spherisorb S5W silica (5 pm) 25 cm X 10 mm column in a Spectra-Physics Model 8100 gradient liquid chromatograph interfaced to a Model 8400 UV/visible detector (A = 254 nm) and Isco fraction collector. The solvent program started with 100% hexane for 10 min, followed by a 5-min linear gradient to 95% hexane/5% CH2C12. The solvent was programmed over the next 25 min to 100% CH2C12,where it was held for 10 min, then programmed to 100% acetonitrile over 10 min, held isocratic for 10 min, and then programmed back to the initial conditions. The flow rate was 3 mL min-l. After 1min, fractions were collected every 9 min, and fraction 6, which contained the nitrodibenzopyranones, was collected in higher resolution subfractions. The quantification of the nitrodibenzopyranone isomers was achieved in the subfraction of fraction 6 covering the retention time range of 56-63 min. This fraction was concentrated to neardryness under a flow of dry nitrogen and then dissolved in 50 pL of dichloromethane for analysis by combined gas chromatography/mass spectrometry (GC/MS). The NIST SRM 1650 diesel particulate material was Soxhlet-extracted in 200-mg aliquots with 200 mL of dichloromethane. The 3-nitrodibenzopyranone internal standard was spiked onto the thimble wall prior to the extraction. The extracts were filtered using 0.2-pm Acrodisc Teflon filter disks, concentrated to -0.5-mL volume, and prefractionated on Sep-Pak silica cartridges (Millipore) by eluting with 10 mL of hexane followed by dichloromethane. The dichloromethane eluate contained the nitrodibenzopyranone isomers and was fractionated by normal-phase HPLC as described above. The nitrodibenzopyranone-containingfraction was further separated by reversed-phase HPLC on a Beckman Ultrasphere ODS 5-pm column (1cm X 25 cm) with a Beckman Model 334 gradient liquid chromatograph equipped with a Model 164 UV detector (A = 254 nm). The solvent used was 100% I methanol, and the fraction eluting from 4 to 6 rnin was collected, concentrated, and analyzed by GC/MS as described below. The GC/MS analyses were carried out using a Hewlett-Packard (HP) 5890 GC interfaced to an H P 5971A
-
2208
Envlron. Scl. Technol., Vol. 26, No. 11, 1992
mas8 selective detector
(MSD) operating in the electron impact ionization mode. The column used was a 60 m x 0.25 mm DB-1701, film thickness 0.25 pm (J&W Scientific). Injections were made on-column at 40 "C oven and injector temperature, and after 2 min the oven was programmed at a rate of 25 "C m i d to 200 "C and then at 3 "Cmin-l to 300 OC. During sample injection, the column head pressure was reduced to 20 psi and then programmed at a rate of 50 psi m i d to 45 psi. Detection was made in the selective ion monitoring mode (SIM), measuring the molecular ion (m/z = 241) and major fragment ions at m/z = 183 and 139 (28). 2-Nitrodibenzopyranone and 4nitrodibenzopyranone eluted with retention indexes (RI) of 404.8 and 430.5, respectively [the retention indexes were determined using chrysene (RI = 400) and benzo[e]pyrene (RI = 452.3) as bracketing standards (28)]. The 3-nitrodibenzopyranone internal standard eluted at RI = 415.2. Quantification was made using the integrated m / z = 241 signal, and the relative response factors of the 2- and 3nitrodibenzopyranones were determined from analyses of calibration solutions containing these isomers in varying ratios and at absolute concentrations comparable to those in the samples. The quantification of 4-nitrodibenzopyranone was carried out by assuming the same response factor as for 2-nitrodibenzopyranone. For control blanks, precleaned TIGF filters and PUF plugs were analyzed, and for the diesel analyses, an empty thimble was extracted and carried through the entire sample analysis procedure. In all cases no signals with sufficient intensity for peak integration were observed. Additionally, a fiiter sample which had been extracted and HPLC-fractionated was refractionated using the same HPLC conditions. It was confirmed that no significant change in the distribution of the nitrodibenzopyranones occurred during the HPLC fractionation. Environmental Chamber Investigation of OH Radical-Initiated Reactions of Phenanthrene, 6 8 Dibenzopyran-6-one, and 2,2'-Diformylbiphenyl. A series of experiments were conducted at 298 f 2 K and in the presence of 740 Torr total pressure of dry (300 nm (10, 11, 13-15, 20, 21),and NO was also added to the reactant mixtures. The initial CH30N0 and NO concentrations were in the range (5-25) X 1013and (2.0-24) X 1013molecule cm+, respectively. Irradiations were carried out at the maximun light intensity for 1-10 min, and after the irradiation, a large-volume ( 1ooO-39OO L) gas sample was collected over 1-4 min onto two PUF plugs held in series in a glass holder. Nitric oxide was monitored by a chemiluminescence NO-NO, monitor, and gas-phase phenanthrene concentrations before and after the reactions were
-
measured by GC with flame ionization detection (FID) by collecting 100-cm3gas samples onto Tenax solid adsorbent with thermal desorption at -300 "C onto a 15-m DB-5 megabore column at 60 "C, which was then temperatureprogrammed at 8 "C min-l. The PUF plugs were analyzed for the nitrodibenzopyranones and structurally-related compounds by GC/MS employing the full-scanning mode for compound identification and SIM for quantifications of the nitrodibenzopyranones, using either the H P 5890 GC and H P 5971A MSD system (see above) or a HP 5890 GC and H P 5970 MSD system with a 60 m X 0.248 mm DB-5 column (J&W Scientific) temperature-programmed from 50 to 150 "C at 25 "C rnin-l and then to 325 "C at 4 "C mi& (with helium as the carrier gas). Bioassay. As reported elsewhere, Salmonella typhimurium, strain TA98, without S9 activation was used to assay the HPLC fractions of ambient air samples collected in Claremont, CA (20,24). To enhance sensitivity, a microsuspension preincubation modification of the Ames assay (22,23),which employs a 90-min preincubation of the test substance with elevated test cell densities in a small-volume buffered-saline suspension, was used. Full details of these tests have been given previously (20,24). The activities of 2-nitro-GH-dibenzo[b,d]pyran-6-one, as received and after HPLC purification, were the same within the reproducibility of the assay. Chemicals. 2-Nitro-GH-dibenzo[b,d]pyran-6-0ne, 3nitro-6H-dibenzo[b,d]pyran-6-one (98% stated purity), and phenanthrene (98% stated purity) were obtained from the Aldrich Chemical Co., and 4-nitro-GH-dibenzo[b,d]pyran6-one was synthesized as described previously (28). 6HDibenzo[b,d]pyranB-one was synthesized by reacting diphenic acid (Aldrich Chemical Co.) with hydrogen peroxide and sulfuric acid according to a procedure given by Ota and Okazaki (30). 2,2'-Diformylbiphenyl was obtained from the ozonolysis of phenanthrene in methanol solution (31). The identities and purities of the synthesized compounds were established by GC/MS and high-resolution MS. Results Analyses of Ambient Air Samples and the SRMs. 2-Nitrodibenzopyranone was observed in all of the southern California ambient air filter and PUF plug samples analyzed, and 4-nitrodibenzopyranonewas observed in the majority of these samples. Both isomers were observed in the SRM 1649 urban dust. The results of the quantitative analyses of these southern California air samples are listed in Table I, and the results of analyses of the SRM 1649 urban dust collected in Washington, DC are given in Table 11. 2-Nitrodibenzopyranone was quantified in all of the ambient air samples analyzed, at concentrations ranging between 0.04 and 0.8 ng m-3. In certain samples the GC/MS ion signal for 4-nitrodibenzopyranone was insufficient for peak integration, and hence upper limits equal to the estimated detection limit of -0.02 ng m9 are given. Our data (Table I) show that 2-nitrodibenzopyranone was always more abundant than the 4-isomer in the ambient air particulate samples analyzed, generally by a factor of -2. The abundance ratio of 2-nitrodibenzopyranone /4-nitrodibenzopyranone of -2 in these ambient air samples is, however, different than we observed in samples collected from environmental chamber OH radical-initiated reactions of phenanthrene, where the 2-nitrodibenzopyranone/4-nitrodibenzopyranoneratio was -0.5 (see ref 21 and Table 111). The data in Table I also show that 2- and 4-nitrodibenzopyranones are distributed between the gas and
Table I. Concentrations of 2- and 4-Nitrodibenzopyranones (NDBP) in Ambient Air Particulate- (Filter) and Gas-Phase (PUF) Samples Collected in Claremont, CA (1987), Long Beach, CA (1987), and Riverside, CA (1991)' samplingb date time
filter 2-NDBP 4-NDBP Claremont 0.24 0.21 0.18
Aug 27 Aug28 Aug29
D D D
0.41 0.37 0.24
June 19 June 19 June 24 June25 July 13 July 14 Nov 11 Nov 13O Dec 2 Dec 10
D N D N D N N D N N
0.16 0.04 0.20
Sept 17 Sept 18 Sept 18 Sept 19 Sept 19/20 Oct 14 Oct 15 Oct 16 Oct 17 Oct 17 Oct 18
N D N D N/W N D N D N D
0.29 0.30 0.15 0.30 0.25 0.36
Riversidd 0.19 0.20 0.08 0.14 0.15 0.15
0.32 0.26
0.10 0.11
0.ah
0.21
PUF 2-NDBP 4-NDBP 0.14 0.14'
0.03 0.02c
0.12 0.10 0.03 0.06 0.02 0.07
